Sunday, March 11, 2007
Evolution as Design
RichardDawkins.net has posted a collection of interviews with Daniel Dennett where he responds to questions about evolution. You can seem them at [Understanding Genetics - Daniel Dennett Interview].
I have a bit of a problem thinking of Daniel Dennett as a good genetics teacher. I also have difficulty thinking of him as a good teacher of evolution since he seems to think that natural selection is the only game in town. He is a classic Ultra-Darwinian.
One of the problems with the Dennett approach to evolution is the emphasis on "design" in nature. I disagree with the basic concept that nature looks designed and I certainly disagree with the idea that natural selection is responsible for everything we see around us. Listen to how Dennett describes his view of the biological world.
Do you agree with this perspective? I don't. Not only do I think it's wrong, I think it concedes an important point to Paley and the Intelligent Design Creationists. I favor Evolution by Accident. The biological world doesn't look all that well designed to me.
Subscribe to:
Post Comments
(
Atom
)
27 comments :
I would agree that there is an appearance of design in parts of nature. The human eye is an obvious example because its structure resembles that of a digital camera.
Whether or not the eye is well-designed is a separate issue. From the perspective of a human designer there are obvious flaws but you could argue that, from the perspective of an alien designer who was limited to working with organic materials that had to be grown and bred selectively over a number of generations, the human eye is a reasonably successful outcome.
The problem with the argument from design is that ignores the possibility that its appearance may simply be an artefact of the properties of the Universe itself.
A 'device' that can gather and focus image-bearing light on a 'transducer' which can convert that light into electrical signals which can be processed subsequently so as to reconstitute that image, will have certain properties dictated by the properties of light and the materials which interact with it. In other words, an eye and a digital camera will have similar properties because they perform a similar function. The fact that one is the product of human ingenuity does not imply that the other must be, particularly if we can argue that the other could be the inevitable result of the outworking of a process of n*t*r*l s*l*ct**n.
I'm not able to view the Dennett piece until I'm at home, so I can't comment on that. What I did want to comment on, and thank you for, was the link to the excellent Evolution by Accident page. I very much like its clear explication of deep concepts, and by coincidence (chance?) it also brings into focus some ideas about which I've been thinking quite a bit lately.
It seems to me that much of ID's intuitive appeal (that is, for those to whom it does intuitively appeal) comes from the erroneous "monkeys typing Shakespeare" view of probability. This error arises fundamentally from thinking of evolution as a *goal-directed* activity, whether the goal is seen as near-perfect adaptation to a local environment, or as the creation of humans as the "paragon of animals." But *any* "history...that includes a huge number of chance and random events all of which are contingent upon everything that preceeded them" becomes vanishingly improbable when conceptualized as the path to a particular goal. It is child's play to "prove," using this erroneous view, that you cannot possibly be sitting here reading this. Your existence depends upon the chance that your parents met each other, out of billions of people of the opposite sex on Earth at the time, multiplied by comparably long odds back through your lineage over each of tens of thousands of generations.
Only when we accurately assess what evolution has produced ("I don't see perfection around me...") - from the relatively trivial (e.g., guys have nipples), to the more significant (backaches), to the tragic (cancer, murder, war) - can we begin to think of the process and the associated probabilities correctly: Not as the path to a particular goal, but as the way the movie played out *this time*.
No, I don't agree with delusions of design.
Sometimes there are constraints that makes life's mechanisms looks like ours (for example eyes vs cameras), sometimes the degree of freedom makes them hugely different (for example legs vs wheels).
But the bottom line is that there are neither a possibility to describe all evolved mechanisms in a finite description nor a possibility to predict what will evolve. Specifically, we can't even describe all structures even with a single conflating complexity measure by principle. (According to Murray Gell-Mann, IIRC.) Both problems shows that design isn't describable.
Design theorists are barking up the wrong tree - the only structures that evolution really predicts as I understand it are phylogenetic trees. (Perhaps some mechanism for inheritance is predicted as well, but nothing specific.)
"But *any* "history...that includes a huge number of chance and random events all of which are contingent upon everything that preceeded them" becomes vanishingly improbable when conceptualized as the path to a particular goal."
The difference between a priori probability and a posteriori outcome is a recurrent problem with creationists. You can multiply with the vanishingly small probability that a certain sperm fertilized an ova, and it seems impossible a child resulted at all. (Which I believe is close to the truth, since AFAIK most attempts fail.) But I have never heard a creationist make the obvious conclusion about their mistaken methods of probability from that.
On second thought, perhaps populations and their structures (age, and topologies like ring species) qualify as predicted or at least described structures as well.
I have heard that Dennett is privately even more orthodox genocentric, super-deterministic neo-Darwinian than he lets out publically. Hard to imagine how an one be even more so than Dennett already is in his books and articles!
Someone once said that if you rewind the Tape of Life over and over again, Dennett believes that every time evolution will produce Daniel Dennett writing "Darwin's Dangerous Idea" (IMHO a God-awful book I call "Dennett's Dangerously Stupid Idea").
Larry,
I definitely agree with your misgiving about nature looking designed. To me design is always intentionally forward looking: a designer looks to desired end and uses that to impose structure on the design. I see nothing following that model in the natural world - not galaxies, not Earthly storm systems, and not life. I get especially dismayed with Dennett's life-looks-designed language because it seems he has intentionally tossed creationists a bone.
On the other hand, I think you are distorting Dawkins position. Yes, he does strongly emphasize natural selection as an important evolutionary mechanism, but he does not discount the roles played by other evolutionary mechanisms in accounting for extant diversity, as suggested on your "Evolution by Accident" site. Dawkins acknowledges genetic drift and other forms of catastrophism as being important evolutionary processes. During his recent book tour in the US, he often pointed to gene duplication as another important factor in evolution.
Larry, when you say things like "According to Dawkins, evolution cannot be haphazard, random, or accidental, else it could never achieve the design attributed to it," you risk losing credibility by apparently intentionally mischaracterizing the stance taken by one of your evolutionary colleagues to drive home your point.
Similarly, when you criticize with, "He frequently uses metaphors like 'Climbing Mt. Improbable' that suggest a single goal for a species," you have missed the entire purpose of his metaphor: to counter the creationist argument from improbability by explicating how one evolutionary mechanism, natural selection, works by building on existing structures in adaptationally-sound increments. He does not suggest that a species lies in wait to spring upon the next available adaptation that will move it up some predetermined teleological mountain.
Larry, I agree with you wholeheartedly that many random processes have always played a part in how evolution proceeds, and, therefore, what has constituted the snapshot of biodiversity at any point in time. However, I was really surprised about was how you misrepresent other respected popularizers and researchers in evolution to make your point.
Yes, selection is constrained in many ways.
Sure, drift happens, especially in small populations and/or when selection is weak.
Pleiotropy. Polygenism. Trade-offs. All of that.
Still: Adaptation happens, and natural selection is the only known mechanism. I am sure that Dr. Moran is familiar with many excellent examples from comparative biochemistry (see Hochachka & Somero).
So by all means get rid of the bathwater, but let's keep the baby too.
russ says,
Yes, he does strongly emphasize natural selection as an important evolutionary mechanism, but he does not discount the roles played by other evolutionary mechanisms in accounting for extant diversity, as suggested on your "Evolution by Accident" site. Dawkins acknowledges genetic drift and other forms of catastrophism as being important evolutionary processes. During his recent book tour in the US, he often pointed to gene duplication as another important factor in evolution.
Actually Dawkins does not acknowledge genetic drift (to take just one example) as an important evolutionary process. He tends to look at random genetic drift as just background noise that plays no role at all in evolution at the macroscopic level.
He's quite clear about this in his writing. As long as a phenotype is visible he attributes it to natural selection—as long as he's not challenged. When challenged he will admit to some role for drift, what else can he do? But it doesn't come naturally to him.
Russ says,
Larry, when you say things like "According to Dawkins, evolution cannot be haphazard, random, or accidental, else it could never achieve the design attributed to it," you risk losing credibility by apparently intentionally mischaracterizing the stance taken by one of your evolutionary colleagues to drive home your point.
It's pretty hard to misinterpret an entire book like The Blind Watchmaker. Perhaps you can explain passages like this one at the beginning of Chapter 3?
We have seen that living things are too improbable and too beautifully 'designed' to have come ino existence by chance.
You should also re-read Chapter 2 Good Design to remind yourself of the Dawkins position.
Dawkins' entire point is based on the fact that natural selection is the blind watchmaker, leading to well-designed watches even though that wasn't the purpose in the beginning. There's no room in that philosophy for sloppy watches that don't keep time or, heaven forbid, ones that don't work at all.
"The biological world doesn't look all that well designed to me."
Then it naturally follows that your own reasoning must not be that well designed either :)
We have seen that living things are too improbable and too beautifully 'designed' to have come ino existence by chance.
I don't have my copy of TBW next to me for reference, but methinks those quotation marks are there for a reason.
Actually Dawkins does not acknowledge genetic drift (to take just one example) as an important evolutionary process. He tends to look at random genetic drift as just background noise that plays no role at all in evolution at the macroscopic level.
I was guided to this post by Russ, and being embroiled in work right now I cannot attend to or address all of the points here - I am a relatively new reader of Dawkins and don't know the ins and outs of his evolutionary model yet (and I have not yet read Improbable), but I did remember that in The Extended Phenotype he says something about Wright suggesting that a subtle combination of selection and genetic drift could produce results more adaptive than selection alone, and I don't think Dawkins disputes this.
Larry, I just read Evolution by Accident, and I must agree with Russ - I truly think you're distorting Dawkins' position. You raise a couple good points of disagreement but I believe they are due to Dawkins' emphasis (due to whatever creationist audience he has), not his denial of any evolutionary facts. Do you really he would disagree with what you've written? I didn't disagree with it. Do you really think he eschews the concept of random mutation? Again, I have read neither Improbable nor Watchmaker yet, so perhaps I am wrong here, but in Phenotype Dawkins says: "No matter how strong a potential selection pressure may be, no evolution will result unless there is a genetic variation for it to work on....Lewontin is undoubtedly right that biologists interested in adaptation cannot afford to ignore the question of mutational variation." He goes on to say, however, that complex adaptations rely upon natural selection.
Actually Dawkins does not acknowledge genetic drift (to take just one example) as an important evolutionary process. He tends to look at random genetic drift as just background noise that plays no role at all in evolution at the macroscopic level
on a phenotypic level (as opposed to the DNA level), the default assumption from most geneticists seems to be that most variation is under selection. If something is "visible" to selection, why wouldn't it be worked on?
Even mRNA levels (the closest thing to DNA, so you might their evolution to be dominated by drift) show overwhelming evidence for being under strong selective constraint.
http://www.nature.com/ng/journal/v37/n5/abs/ng1554.html
http://www.nature.com/nature/journal/v438/n7065/abs/nature04114.html
including in humans
http://www.nature.com/nature/journal/v440/n7081/abs/nature04559.html
apologies for the cut-off links. Link for evolution of gene expression in humans, Drosophila, and c elegans
Kristine says,
Do you really he would disagree with what you've written?
Yes, of course he does. We've corresponded extensively about the content of that essay and we even talked about it briefly when I was at his house in October. As a matter of fact, I modified some of the language to more acurately reflect Dawkins' opinion in response to some of his comments.
The debate between Ultras and Pluralists has been going on for thirty years. It's not just someone's overactive imagination as you seem to imply.
Do you really think he eschews the concept of random mutation?
No, of course he doesn't. But that's not the point. The point is whether evolution is limited (or guided) by mutation or whether there's always such an abundance of variation that natural selection will always find the best path. Dawkins believes the latter. It permeates his writing in spite of the fact that he occasionally pays lip service to the idea that mutations aren't always there.
There is no doubt in Dawkins mind that organisms will always get to the top of Mt. Improbable.
Cody says,
I don't have my copy of TBW next to me for reference, but methinks those quotation marks are there for a reason.
Let me help out. Look at the opening sentences to the chapter titled Good Design (Chapter 2).
Natural selection is the blind watchmaker, blind because it does not see ahead, does not plane consequences, has no purpose in view. Yet the living results of natural selection overwhelmingly impress us with the apearance of design, as if by a master watchmaker, impress us with the illusion of design and planning. The purpose of this book is to resolve this paradox to the satisfaction of the reader, and the purpose of this chapter is further to impress the reader with the power of the illusion of design. We shall look at a particular example and shall conclude that, when it comes to complexity and beauty of design, Paley hardly even began to state the case.
The difference between the adaptionists like Dawkins and Dennett and the pluralists like Gould and Lewontin is that the pluralists are not so mesmerized by the illusion of design.
When we look at the organization of the mammalian genome, for example, we don't see the same hand of the watchmaker that Dawkins sees. It's a question of emphasis. Dawkins sees design everywhere in nature and I only see it in some places. Lots of evolution looks sloppy and accidental to me.
My "watchmaker" acts more like Rube Goldberg or the tinkerer of François Jacob.
ccp says,
Yes, selection is constrained in many ways. Sure, drift happens, especially in small populations and/or when selection is weak.
This is exactly the perception that I'm fighting. You are relegating drift to a minor role that's only important in small populations or when natural selection isn't working. That's wrong. It's what Gould and Lewontin criticized in their famous "Spandrels" paper of 1979. They said,
At this point, some evolutionists will protest that we are caricaturing their view of adaptation. After all, do they not admit genetic drfit, allometry, and a variety of reasons for non-adaptive evolution? They do, to be sure, but we make a diferent point. In natural history, all possible things happen sometimes; you generally do not support your favourite phenomenon by declaring rivals impossible in theory. Rather, you acknowldege the rival but circumscribe its domain of action so narrowly that it cannot have any importance in the affairs of nature. Then, you congratulate yourself for being such an undogmatic and ecumenical chap.
ccp continues,
Still: Adaptation happens, and natural selection is the only known mechanism. I am sure that Dr. Moran is familiar with many excellent examples from comparative biochemistry (see Hochachka & Somero).
Of course adaptation happens. That's the view of the pluralists. Lots of different mechanisms give rise to the results we see in the history of life. Nothing is excluded, especially not natural selection.
But there's a lot more to evolution than just natural selection and adaptation.
So by all means get rid of the bathwater, but let's keep the baby too.
I couldn't agree more. The Ultra-Darwinians, with their myopic focus on adaptation, have chucked out the genetic drift baby and lots of other important considerations in evolutionary theory. The pluralists, like myself, are trying to put them back in the bathwater. As a matter of fact, we don't even throw out the bathwater!
p-ter says,
on a phenotypic level (as opposed to the DNA level), the default assumption from most geneticists seems to be that most variation is under selection.
Some so-called "geneticists" behave like that but they're wrong to do so. They begin by assuming the very mechanism they're supposed to be proving.
The correct default assumption is no positive mechanism of evolution; i.e., random genetic drift. The way most geneticists behave is to look for deviations from the default assumption of no selection. Often it's by looking at the Hardy-Weinberg distribution to see if there's anything to suggest selection.
As it happens, I'm reading a paper on the detection of natural selection in humans (Harris & Meyer, 2006 The Molecular Signal of Selection Underlying Human Adaptations). It's an extensive review that I recommend very highly to anyone who is interested in this topic.
The authors are interested in the ways to detect selection and the review focuses very much on methodology. Here's an excerpt from the introduction,
Notice that the study of natural selection using genetic data is based upon tests of the null hypothesis of neutrality, rather than tests of natural selection. Neutral evolution is an analytically tractable model of evolution, which makes simple predictions about the frequencies of alleles and polymorphisms, the expected proportions of polymorphism, and divergence for different genes of classes of mutaiton (Kimura, 1983). Thus, the tests employed in the study of natural selection upon genetic variation are more adequately defined as tests of the null hypothesis of neutrality or neutrality tests.
This paper is about genetic tests—those where a gene is known or strongly suspected. You were referring to phenotypic characters where the underlying genetic component may not be known. But that doesn't make a difference. The default assumption should still be neutrality and genetic drift. You need to provide evidence for adaptation before assuming it. That's why "just-so" stories have such a bad reputation among evolutionary biologists.
BTW, the idea that most variation in a population is under selection has been decisively disproven. I hope you didn't mean to imply otherwise.
You were referring to phenotypic characters where the underlying genetic component may not be known. But that doesn't make a difference. The default assumption should still be neutrality and genetic drift. You need to provide evidence for adaptation before assuming it.
yes, and that's what those papers I cited did. they rejected neutrality for the evolution of gene expression traits.
The debate between Ultras and Pluralists has been going on for thirty years. It's not just someone's overactive imagination as you seem to imply.
I did not imply or even think that at all and I'm pretty new to that debate. At any rate I'm sorry that I said anything.
I'm curious about the anti-Design viewpoint expressed here. It seems to take an idealized view of design as achieving some sort of effortless perfection. Real design is much closer to tinkering than that -- in both cases it involves generating possibilities and trying them out. A good designer might be able to evaluate a possibility in their head, or in a computer, rather than by actually building it, but the generate-and-test cycle is still integral.
This is why technology develops incrementally through its own process of evolution -- not natural selection, but not entirely different either.
Here's a book that explores this theme.
I'm not a party to the great evolutionary debates, but to me it seems obvious that natural systems seem to be like any designed technological system -- that is to say, they look designed but not as if by a perfect designer, rather by a designer who is constrained by history and the requirement to make incremental improvements. Your computer keyboard was designed but like a naturally evolved system it bears the marks of its history (in this case, a less-than-ideal keyboard layout).
Torbjörn Larsson said...
No, I don't agree with delusions of design.
Sometimes there are constraints that makes life's mechanisms looks like ours (for example eyes vs cameras), sometimes the degree of freedom makes them hugely different (for example legs vs wheels).
"Legs v wheels" is an interesting idea. Could an animal evolve wheels?
"Why don't any animals have wheels?" is one of my favorite see-if-they-can-think-on-their-feet graduate oral exam questions.
While I'm posting here, I'd like to point out that, in his reply to my previous comment on this thread, Dr. Moran changed my "drift happens, especially in small populations and/or when selection is weak," an entirely accurate and, as far as I know, uncontroversial statement, to "You are relegating drift to a minor role that's only important in small populations or when natural selection isn't working."...which is a very fibrous straw man.
I'm a pluralist too! But I'm a pluralist who acknowledges the ubiquity of adaptation (and, hence, selection--I do not subscribe to "current utility alone" definitions of adaptation in the comparative physiology of the animals I study. Evidence of selection constrained is also common and interesting.
oops. Close paren after the last "adaptation" in the previous comment.
CCP,
While I'm posting here, I'd like to point out that, in his reply to my previous comment on this thread, Dr. Moran changed my "drift happens, especially in small populations and/or when selection is weak," an entirely accurate and, as far as I know, uncontroversial statement, to "You are relegating drift to a minor role that's only important in small populations or when natural selection isn't working."...which is a very fibrous straw man.
Point taken. I didn't mean to misinterpret your statement. But it still sounds restrictive to me. Here's how I would put it. Let's see if you agree.
Random genetic drift happens in populations of all sizes at all times. It interacts with selection when an allele is beneficial or deleterious causing most beneficial alleles to be lost and causing some deleterious alleles to be fixed. Random genetic drift is the only game in town when an allele is neutral or nearly neutral. The vast majority of all alleles that are fixed in a population are fixed by random genetic drift and not by natural selection.
Is that what you meant?
The point about "small populations" is important. Most people try to dismiss drift as something that's only important in small populations. This isn't true. The net effect of random genetic drift does not depend on population size. While it's true that for any given mutation the effect of drift depends on the size of the population this is offset in real populations by the increased mutation rate in large popualtions.
Perhaps that's what you meant to say but you were just trying to conserve words. If so, I apologize for assuming that you were spouting the usual myths about drift. I urge you to make your point more clearly in the future.
Well, the truth is I'm now out of my depth...I have not dealt in population genetics above the teaching-intro-to-bio level for a couple of decades.
I did find my eyebrows going up at a couple of your statements though:
"The vast majority of all alleles that are fixed in a population are fixed by random genetic drift and not by natural selection."
Really? Alleles with phenotypic effect at the amino-acid level? At the functional protein level?
"The net effect of random genetic drift does not depend on population size. While it's true that for any given mutation the effect of drift depends on the size of the population this is offset in real populations by the increased mutation rate in large popualtions."
I find that surprising too. I thought drift was essentially the result of sampling error. Surely fixation is more likely and more rapid in smaller populations? Much more likely to get all tails in 5 coinflips than in 50 and all that, and this must be what you mean by the effect of drift on any individual mutation.
But I am having trouble conceptualizing the relationships between mutation rate, drift, and population size...the frequency of any particular neutral allele can either increase or decrease due to drift but it's not clear to me why an allele's population frequency should depend one way or the other on population size.
I don't claim to know, I'm really asking about this.
Post a Comment